The AI-wealth premium moves from the model to the pipes
A $250 million check from Dave Duffield and a $15 million round anchored to 170 custodian connections say the rail is now the scarce asset in wealth technology.
Dave Duffield wrote the entire $250 million check, placing Ridgeline's Series E at a $1.425 billion valuation and answering, by his own hand, the question the wealth-technology market has been circling: what is the artificial-intelligence trade in advice actually paying for? PWD's tracking shows the round closed Sept. 16 with two parties, Ridgeline and Duffield, and no outside lead—a founder funding his own expansion sets the price without answering to a partner's growth chart, and he accepts that the round carries exactly one mark.
The valuation rests on an argument: Ridgeline's $750 billion of committed platform assets is a pipeline its two named clients do not fill. The round reads as Duffield buying "the RIA's plumbing," and commitments at that scale pile up where switching costs live—inside the connections an advisory book already runs through. A portfolio engine can be rebuilt by anyone with a competent team; the connections took years and counterparties to assemble.
Pave makes the same case with a smaller check: the AI portfolio platform raised a $15 million Series A at a $100 million pre-money valuation, and the asset that anchored that price was not the model but 170 custodian connections. Divide the valuation by the connections and each implies roughly $590,000 of company value, a strange price for software unless the connection is what is being bought, because model quality is comparable across vendors and improving for all of them while a custodian integration is a contract with a counterparty on the other side of it. Pave just put a number on the rail.
Envestnet reached the same conclusion from the buy side: its deal for Vestmark reads as a financing story in product clothing, and the tell is the "much larger war chest" the $8 trillion platform is spending out of. Integration capacity at that scale is a multi-year build done internally, and the buyer with the balance sheet to skip the queue generally does—what the deal buys, on that reading, is the asset Pave is selling: the position between an advisor and the systems the money moves through.
Three financings converging on the same verdict is a pattern worth marking: in maturing software categories the margin migrates to whichever layer is hardest to replicate, and in advice that layer is the rail—the data feeds, the custodian handshakes, the compliance sign-offs, and the distribution that gets a tool in front of advisors who already have too many tools. Model capability is converging and cheap; a custodian connection resists both. The three prices encode a common judgment, and it is the correct one—what looked like the product is now table stakes, and the platform holding the most connections decides which features its advisors ever see.
The rest of the tape
Tuesday's Form D filings put the smaller end in view: CAZ Founders Class - GPG Fund Investment, L.P. reported $175.3 million sold against an undisclosed offering, with Christopher Zook among the related persons named; Dyme Capital sold its $3.3 million private equity offering to the dollar; and D. Boral's Series XII Atoms vehicle moved $2.4 million of a $3 million raise, while two venture series listing Belltower Fund Group among their related persons came in at $290,000 and $18,000—sizes that suggest single-position vehicles rather than franchises.
Pave's implied $590,000 per custodian connection is the number to watch. If the next wealth-technology round prices connections the same way, the plumbing has become the asset class and the model vendors are selling into a market that no longer pays for differentiation. If the next round is won on benchmark performance instead, this week's three prices were a coincidence.